>I will forgive him his SYS1.NUCLEUS jibe, since the rest of his post >makes it clear that he knows very well where the real problem lies.
Maybe I don't know. The only problem that I know of, and the reason that I created deferred LPA wait,was the question of how an application, starting after IPL, could know for sure that modules it needed to be in LPA (and that needed to come from a PDSE) were actually available to the application. Obviously for normal LPA, this is known because the IPL would not have gotten to the point of starting applications without normal LPA having been built. In practice, no one had complained about this, but it had bothered me, so perhaps the silent majority had been too silent and left it up to us to do what we thought needed doing. >RAS problems with them at what I shall now call LPA-construction >time needed to be addressed. I'm not sure, but perhaps John misconstrued what I was getting at. Having to have all the PDSE support code in common storage (and its data in the master address space) rather than in a separate address space is a RAS problem (and possibly a VSCR problem) that is not solvable without a massive restructure to allow multiple address spaces before LPA is built. That won't happen.. Yeah, you could conceive of the code living temporarily in master's private and then the "real copy" live in its own address space later, but that isn't going to happen either. This simply comes back to the requirement: there is none. I was not referring to RAS problems with PDSEs which would be,an entirely different matter and would not need an LPA discussion to be brought forward. >I want to get rid of PDSs. Want's are nice, but do not necessarily represent requirements. There is no reason why a lot of money should be spent to make this happen. Peter Relson z/OS Core Technology Design ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
